def test_deprecated_print_cyclic():
p = Permutation(0, 1, 2)
try:
Permutation.print_cyclic = True
with warns_deprecated_sympy():
assert sstr(p) == '(0 1 2)'
with warns_deprecated_sympy():
assert srepr(p) == 'Permutation(0, 1, 2)'
with warns_deprecated_sympy():
assert pretty(p) == '(0 1 2)'
with warns_deprecated_sympy():
assert latex(p) == r'\left( 0\; 1\; 2\right)'
Permutation.print_cyclic = False
with warns_deprecated_sympy():
assert sstr(p) == 'Permutation([1, 2, 0])'
with warns_deprecated_sympy():
assert srepr(p) == 'Permutation([1, 2, 0])'
with warns_deprecated_sympy():
assert pretty(p, use_unicode=False) == '/0 1 2\\\n\\1 2 0/'
with warns_deprecated_sympy():
assert latex(p) == \
r'\begin{pmatrix} 0 & 1 & 2 \\ 1 & 2 & 0 \end{pmatrix}'
finally:
Permutation.print_cyclic = None
def pretty_config(self, tex):
start = datetime.now()
import sympy
from sympy.parsing.sympy_parser import parse_expr, standard_transformations, implicit_multiplication_application
from sympy import init_printing
from sympy.printing import pretty
from sympy import Eq, Symbol
# definition of the transformation used in parse_expr
transformations = transformations = (
standard_transformations + (implicit_multiplication_application, ))
# ^ is considered the bitwise-XOR in programming, as such, ** is adopted in python.
tex = str(tex).replace("^", "**")
# Sympy does not handle equations well ,so we need to split them
if ("=" in tex):
rhs, lhs = tex.split('=')
try:
rhs = parse_expr(rhs, transformations=transformations)
lhs = parse_expr(lhs, transformations=transformations)
except SyntaxError as e:
print(e)
print(
f"[StepPyStep][parse_tex][pretty_config] runtime: {datetime.now() - start}"
)
return tex
# Actual pretty print of sympy
# using more common unicode that looks bette
rhs = pretty(rhs, use_unicode=True)
rhs = rhs.replace("╲╱", "√")
rhs = rhs.replace("sqrt", "√")
lhs = pretty(lhs, use_unicode=True)
lhs = lhs.replace("╲╱", "√")
lhs = lhs.replace("sqrt", "√")
tex = rhs + "=" + lhs + "\n"
print(
f"[StepPyStep][parse_tex][pretty_config] runtime: {datetime.now() - start}"
)
return tex
# Without equations the procedure is very simple
else:
sympy_expr = parse_expr(tex, transformations=transformations)
print(
f"[StepPyStep][parse_tex][pretty_config] runtime: {datetime.now() - start}"
)
pretty_expr = pretty(sympy_expr)
pretty_expr = pretty_expr.replace("╲╱", "√").replace("sqrt", "√")
return pretty_expr
def _report_about_generated_boolean_mtl_problem(functions, tasks):
"""Log a report about the generated synthetic Boolean MTL problem
represented by the given functions and tasks.
Note: The logger object must be a valid Logger.
Parameters
----------
functions : list
A list of Boolean functions comprised of Boolean operators from
sympy.logic, one function for each task group.
tasks : list
Either a list of Bunch objects corresponding to Boolean function
learning tasks,
or a list of lists of Bunch objects, where each list corresponds
to a set of different learning sets for each task.
"""
# extract group names from tasks' ids
group_names = []
for tl in tasks:
if isinstance(tl, list):
t = tl[0]
else:
t = tl
match = re.search(r"(Group \d+)", t.ID)
group_name = match.group(1)
if group_name not in group_names:
group_names.append(group_name)
if len(group_names) != len(functions):
raise ValueError("The number of task groups doesn't correspond to the "
"number of Boolean functions.")
logger.debug("Report about the generated synthetic Boolean MTL problem:")
logger.debug(" Boolean function of each group:")
for group_name, func in zip(group_names, functions):
# NOTE: sympy's pretty() function returns a unicode string, so the
# string literal must also be a unicode string
logger.debug(u" - {}: {}".format(group_name, pretty(func,
wrap_line=False)))
logger.debug(" % of True values in y for each task:")
sum_true = 0
sum_total = 0
for tl in tasks:
if isinstance(tl, list):
for i, t in enumerate(tl):
cur_true = sum(t.target == True)
cur_len = len(t.target)
sum_true += cur_true
sum_total += cur_len
logger.debug(" - {} (learning set #{}): {}".\
format(t.ID, i, cur_true / cur_len))
else:
t = tl
cur_true = sum(t.target == True)
cur_len = len(t.target)
sum_true += cur_true
sum_total += cur_len
logger.debug(" - {}: {}".format(t.ID, cur_true / cur_len))
logger.debug(" Average % of True values in y (across all tasks): {}".\
format(sum_true / sum_total))
def _report_about_generated_boolean_mtl_problem(functions, tasks):
"""Log a report about the generated synthetic Boolean MTL problem
represented by the given functions and tasks.
Note: The logger object must be a valid Logger.
Parameters
----------
functions : list
A list of Boolean functions comprised of Boolean operators from
sympy.logic, one function for each task group.
tasks : list
Either a list of Bunch objects corresponding to Boolean function
learning tasks,
or a list of lists of Bunch objects, where each list corresponds
to a set of different learning sets for each task.
"""
# extract group names from tasks' ids
group_names = []
for tl in tasks:
if isinstance(tl, list):
t = tl[0]
else:
t = tl
match = re.search(r"(Group \d+)", t.ID)
group_name = match.group(1)
if group_name not in group_names:
group_names.append(group_name)
if len(group_names) != len(functions):
raise ValueError("The number of task groups doesn't correspond to the "
"number of Boolean functions.")
logger.debug("Report about the generated synthetic Boolean MTL problem:")
logger.debug(" Boolean function of each group:")
for group_name, func in zip(group_names, functions):
# NOTE: sympy's pretty() function returns a unicode string, so the
# string literal must also be a unicode string
logger.debug(u" - {}: {}".format(group_name,
pretty(func, wrap_line=False)))
logger.debug(" % of True values in y for each task:")
sum_true = 0
sum_total = 0
for tl in tasks:
if isinstance(tl, list):
for i, t in enumerate(tl):
cur_true = sum(t.target == True)
cur_len = len(t.target)
sum_true += cur_true
sum_total += cur_len
logger.debug(" - {} (learning set #{}): {}".\
format(t.ID, i, cur_true / cur_len))
else:
t = tl
cur_true = sum(t.target == True)
cur_len = len(t.target)
sum_true += cur_true
sum_total += cur_len
logger.debug(" - {}: {}".format(t.ID, cur_true / cur_len))
logger.debug(" Average % of True values in y (across all tasks): {}".\
format(sum_true / sum_total))
def test_pretty():
x = symbols('x')
assert pretty(Q.positive(x)) == "Q.positive(x)"
def test_pretty():
assert pretty(Q.positive(x)) == "Q.positive(x)"
def _generate_boolean_data(a, d, n, g, tg, noise, random_seed,
n_learning_sets=1):
"""Generate a synthetic MTL problem of learning Boolean functions according
to the given parameters.
Parameters
----------
a : int
Number of attributes/variables of the generated Boolean functions.
d : int
The expected number of attributes/variables in a disjunct.
n : int
The number of examples for each task to generate.
g : int
The number of task groups to generate. Each task group shares the
same Boolean functions.
tg : int
The number of tasks (with their corresponding data) to generate for
each task group.
noise : float
The proportion of examples of each task that have their class values
determined randomly.
random_seed : int
The random seed with which to initialize a private Random object.
n_learning_sets : int
The number of different learning sets to create for each task.
Returns
-------
tasks : list
If n_learning_sets == 1, a list of Bunch objects corresponding to
Boolean function learning tasks.
Otherwise, a list of lists of Bunch objects, where each list corresponds
to a set of different learning sets for each task.
funcs : list
A list of Boolean functions comprised of Boolean operators from
sympy.logic, one function for each task group.
attr : list of sympy's Symbol objects representing the attributes of the
generated functions
"""
rnd = random.Random(random_seed)
# generate Boolean functions
attrs = []
funcs = []
for i in range(g):
attr, func = generate_boolean_function(a, d,
random_seed=rnd.randint(1, 100))
attrs.append(attr)
funcs.append(func)
# generate examples for all tasks
tasks = [[] for i in range(g * tg)]
for i in range(g):
attr, func = attrs[i], funcs[i]
attr_names = [str(a_) for a_ in attr]
for j in range(tg):
# NOTE: sympy's pretty() function returns a unicode string, so
# the string literal must also be a unicode string
descr = (u"Synthetic boolean data for task {} of group {} "
"(function: {})".format(j + 1, i + 1, pretty(func,
wrap_line=False)))
id = "Group {}, task {}".format(i + 1, j + 1)
for k in range(n_learning_sets):
X, y = generate_examples(attr, func, n, noise=noise,
random_state=rnd.randint(1, 100))
tasks[i * tg + j].append(Bunch(data=X, target=y,
feature_names=attr_names,
DESCR=descr, ID=id))
if n_learning_sets == 1:
tasks = [t[0] for t in tasks]
return tasks, funcs, attr
update_progress(1.* (i + 1) / n_funcs)
print
_report_about_generated_boolean_mtl_problem(funcs, tasks)
return tasks, tasks_complete_test_sets
if __name__ == "__main__":
# generate a Boolean function with 8 variables and disjuncts with an average
# length of 4
a, d = 8, 4
attr, func = generate_boolean_function(a, d, random_seed=2)
# NOTE: sympy's pretty() function returns a unicode string, so the string
# literal must also be a unicode string
print u"Boolean function (a={}, d={}): {}".format(a, d,
pretty(func, wrap_line=False))
X, y = generate_examples(attr, func, n=1000, random_state=10)
print "% of True values in y: {:.2f}".format(100 * sum(y == True) / len(y))
X_noise, y_noise = generate_examples(attr, func, n=1000, noise=0.3,
random_state=10)
# try different learning algorithms in scikit-learn and report their
# cross-validation scores
from sklearn.linear_model import LogisticRegression
from sklearn import cross_validation
lr = LogisticRegression()
print "Log. reg. scores: ", cross_validation.cross_val_score(lr, X, y, cv=5)
from sklearn.tree import DecisionTreeClassifier
dt = DecisionTreeClassifier()
print "Dec. tree scores: ", cross_validation.cross_val_score(dt, X, y, cv=5)
from sklearn.naive_bayes import MultinomialNB
def test_pretty():
x = symbols('x')
assert pretty(Q.positive(x)) == "Q.positive(x)"
X = S1T
Y = S1
reps = [(X[i] * Y[j], int(i == j)) for i in range(3) for j in range(3)]
X = P0T
Y = P0
reps += [(X[i] * Y[j], int(i == j)) for i in range(2) for j in range(2)]
def eval(expr):
return expr.expand().subs(reps)
for v in S1T:
for u in S1:
print('<%s, %s> = %s' % (pretty(v), pretty(u), eval(v * u)))
print()
for v in P0T:
for u in P0:
print('<%s, %s> = %s' % (pretty(v), pretty(u), eval(v * u)))
print()
def transpose(expr):
expr = expr.expand()
expr = expr.subs(zip(S1, symbols('x:3')))
expr = expr.subs(zip(S1T, S1))
expr = expr.subs(zip(symbols('x:3'), S1T))
expr = expr.subs(zip(P0, symbols('x:2')))
expr = expr.subs(zip(P0T, P0))
def test_pretty():
assert pretty(Q.positive(x)) == "Q.positive(x)"
assert pretty(
{Q.positive, Q.integer}) == "{Q.integer, Q.positive}"
def _generate_boolean_data(a,
d,
n,
g,
tg,
noise,
random_seed,
n_learning_sets=1):
"""Generate a synthetic MTL problem of learning Boolean functions according
to the given parameters.
Parameters
----------
a : int
Number of attributes/variables of the generated Boolean functions.
d : int
The expected number of attributes/variables in a disjunct.
n : int
The number of examples for each task to generate.
g : int
The number of task groups to generate. Each task group shares the
same Boolean functions.
tg : int
The number of tasks (with their corresponding data) to generate for
each task group.
noise : float
The proportion of examples of each task that have their class values
determined randomly.
random_seed : int
The random seed with which to initialize a private Random object.
n_learning_sets : int
The number of different learning sets to create for each task.
Returns
-------
tasks : list
If n_learning_sets == 1, a list of Bunch objects corresponding to
Boolean function learning tasks.
Otherwise, a list of lists of Bunch objects, where each list corresponds
to a set of different learning sets for each task.
funcs : list
A list of Boolean functions comprised of Boolean operators from
sympy.logic, one function for each task group.
attr : list of sympy's Symbol objects representing the attributes of the
generated functions
"""
rnd = random.Random(random_seed)
# generate Boolean functions
attrs = []
funcs = []
for i in range(g):
attr, func = generate_boolean_function(a,
d,
random_seed=rnd.randint(1, 100))
attrs.append(attr)
funcs.append(func)
# generate examples for all tasks
tasks = [[] for i in range(g * tg)]
for i in range(g):
attr, func = attrs[i], funcs[i]
attr_names = [str(a_) for a_ in attr]
for j in range(tg):
# NOTE: sympy's pretty() function returns a unicode string, so
# the string literal must also be a unicode string
descr = (u"Synthetic boolean data for task {} of group {} "
"(function: {})".format(j + 1, i + 1,
pretty(func, wrap_line=False)))
id = "Group {}, task {}".format(i + 1, j + 1)
for k in range(n_learning_sets):
X, y = generate_examples(attr,
func,
n,
noise=noise,
random_state=rnd.randint(1, 100))
tasks[i * tg + j].append(
Bunch(data=X,
target=y,
feature_names=attr_names,
DESCR=descr,
ID=id))
if n_learning_sets == 1:
tasks = [t[0] for t in tasks]
return tasks, funcs, attr
update_progress(1. * (i + 1) / n_funcs)
print
_report_about_generated_boolean_mtl_problem(funcs, tasks)
return tasks, tasks_complete_test_sets
if __name__ == "__main__":
# generate a Boolean function with 8 variables and disjuncts with an average
# length of 4
a, d = 8, 4
attr, func = generate_boolean_function(a, d, random_seed=2)
# NOTE: sympy's pretty() function returns a unicode string, so the string
# literal must also be a unicode string
print u"Boolean function (a={}, d={}): {}".format(
a, d, pretty(func, wrap_line=False))
X, y = generate_examples(attr, func, n=1000, random_state=10)
print "% of True values in y: {:.2f}".format(100 * sum(y == True) / len(y))
X_noise, y_noise = generate_examples(attr,
func,
n=1000,
noise=0.3,
random_state=10)
# try different learning algorithms in scikit-learn and report their
# cross-validation scores
from sklearn.linear_model import LogisticRegression
from sklearn import cross_validation
lr = LogisticRegression()
print "Log. reg. scores: ", cross_validation.cross_val_score(lr,
X,
def test_pretty():
assert pretty(Q.positive(x)) == "Q.positive(x)"
assert pretty(set([Q.positive, Q.integer])) == "set([Q.integer, Q.positive])"
def test_pretty():
x = symbols('x')
assert pretty(Assume(x, 'positive')) == "Assume(x, 'positive')"
def test_pretty():
x = symbols('x')
assert pretty(Assume(x, 'positive')) == "Assume(x, 'positive')"
def test_pretty():
assert pretty(Q.positive(x)) == "Q.positive(x)"
def test_pretty():
assert pretty(Q.positive(x)) == "Q.positive(x)"
assert pretty(set([Q.positive,
Q.integer])) == "set([Q.integer, Q.positive])"
def test_pretty():
x = symbols('x')
assert pretty(Assume(x, 'positive', True)) == 'Assume(x, positive, True)'
